[sr-dev] Crash in free_sip_msg -> reset_ruid

Daniel-Constantin Mierla miconda at gmail.com
Wed Oct 2 22:42:18 CEST 2013


Nothing suspect so far. The crash is induced by my commit in to 
reset_ruid() in free sip msg function. It is needed otherwise is a leak, 
as set_ruid() is doing pkg_malloc(). So before was not crashing in this 
free, but leaked memory.

If you can reproduce it tomorrow, can you try to grab the logs with 
debug=3 (eventually with cfgtrace=1 for debugger module)? Apparently is 
an overwrite of the pointer, even should be a simple processing (I 
guess), because it is an OPTIONS request.

Cheers,
Daniel

On 10/2/13 9:54 PM, Hugh Waite wrote:
> usrloc db_mode is 3.
> db_ops_ruid is 1 as well
>
> Hugh
> On 02/10/2013 20:41, Daniel-Constantin Mierla wrote:
>> One more question, what is the value of db_mode for usrloc module?
>>
>> Cheers,
>> Daniel
>>
>> On 10/2/13 9:28 PM, Hugh Waite wrote:
>>> On 02/10/2013 19:18, Daniel-Constantin Mierla wrote:
>>>> Hello,
>>>>
>>>> can you give bt full as well as kamailio -v output? Any log error 
>>>> messages?
>>>>
>>>> Also, it would be good to recompile with MEMDBG=1 and watch for 
>>>> errors in the logs to see if there is a buffer overflow.
>>>>
>>>> Cheers,
>>>> Daniel
>>>>
>>>> On 10/2/13 7:19 PM, Hugh Waite wrote:
>>>>> Hi,
>>>>> We've had some more crashes on the current master build.
>>>>> (gdb) bt
>>>>> #0  qm_insert_free (qm=0x7fc1e1b9e010, p=<value optimized out>) at 
>>>>> mem/q_malloc.c:181
>>>>> #1  qm_free (qm=0x7fc1e1b9e010, p=<value optimized out>) at 
>>>>> mem/q_malloc.c:527
>>>>> #2  0x000000000055027f in reset_ruid (msg=0x7fc1e1c35360) at 
>>>>> parser/msg_parser.c:911
>>>>> #3  free_sip_msg (msg=0x7fc1e1c35360) at parser/msg_parser.c:730
>>>>> #4  0x00000000004a4012 in receive_msg (buf=<value optimized out>, 
>>>>> len=<value optimized out>, rcv_info=<value optimized out>) at 
>>>>> receive.c:297
>>>>> #5  0x000000000052a251 in tcp_read_req (con=0x7fc1ca4c6e00, 
>>>>> bytes_read=0x7fff041b327c, read_flags=0x7fff041b3274) at 
>>>>> tcp_read.c:1387
>>>>> #6  0x000000000052c53b in handle_io (fm=<value optimized out>, 
>>>>> events=1, idx=-1) at tcp_read.c:1617
>>>>> #7  0x000000000052eb69 in io_wait_loop_epoll (unix_sock=<value 
>>>>> optimized out>) at io_wait.h:1092
>>>>> #8  tcp_receive_loop (unix_sock=<value optimized out>) at 
>>>>> tcp_read.c:1728
>>>>> #9  0x00000000004fc0eb in tcp_init_children () at tcp_main.c:4959
>>>>> #10 0x000000000046c3d5 in main_loop () at main.c:1702
>>>>> #11 0x000000000046dec9 in main (argc=<value optimized out>, 
>>>>> argv=<value optimized out>) at main.c:2533
>>>>>
>>>>> (gdb) frame 2
>>>>> #2  0x000000000055027f in reset_ruid (msg=0x7fc1e1c35360) at 
>>>>> parser/msg_parser.c:911
>>>>> 911                     pkg_free(msg->ruid.s);
>>>>> (gdb) p msg->ruid
>>>>> $7 = {s = 0x845d20 "", len = 20}
>>>>>
>>>>> Might this be related to the changes made on Sept 19th to the 
>>>>> free_sip_msg functions?
>>>>>
>>>>> Regards,
>>>>> Hugh
>>>>>
>>>>
>>> Extra output below.
>>> Nothing was printed in the logs (WARNING or ERROR level) before the 
>>> crash. It seemed to be quite reproduceable when there was traffic 
>>> being sent to registered websocket clients, but there is no-one 
>>> online now. We'll have multiple people logged on tomorrow morning.
>>>
>>> Regards,
>>> Hugh
>>>
>>> kamailio -v
>>> version: kamailio 4.1.0-dev9 (x86_64/linux)
>>> flags: STATS: Off, USE_TCP, USE_TLS, TLS_HOOKS, USE_RAW_SOCKS, 
>>> DISABLE_NAGLE, USE_MCAST, DNS_IP_HACK, SHM_MEM, SHM_MMAP, 
>>> PKG_MALLOC, USE_FUTEX, FAST_LOCK-ADAPTIVE_WAIT, USE_DNS_CACHE, 
>>> USE_DNS_FAILOVER, USE_NAPTR, USE_DST_BLACKLIST, HAVE_RESOLV_RES
>>> ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 
>>> 16, MAX_URI_SIZE 1024, BUF_SIZE 65535, DEFAULT PKG_SIZE 4MB
>>> poll method support: poll, epoll_lt, epoll_et, sigio_rt, select.
>>> id: unknown
>>> compiled on 13:35:36 Oct  2 2013 with gcc 4.4.7
>>>
>>> (gdb) bt full
>>> #0  qm_insert_free (qm=0x7fc1e1b9e010, p=<value optimized out>) at 
>>> mem/q_malloc.c:181
>>>         f = 0x845d10
>>>         prev = 0x65332d3231653163
>>>         hash = 0
>>> #1  qm_free (qm=0x7fc1e1b9e010, p=<value optimized out>) at 
>>> mem/q_malloc.c:527
>>>         f = 0x845d10
>>>         size = <value optimized out>
>>>         next = <value optimized out>
>>>         prev = <value optimized out>
>>>         __FUNCTION__ = "qm_free"
>>> #2  0x000000000055027f in reset_ruid (msg=0x7fc1e1c35360) at 
>>> parser/msg_parser.c:911
>>> No locals.
>>> #3  free_sip_msg (msg=0x7fc1e1c35360) at parser/msg_parser.c:730
>>> No locals.
>>> #4  0x00000000004a4012 in receive_msg (buf=<value optimized out>, 
>>> len=<value optimized out>, rcv_info=<value optimized out>) at 
>>> receive.c:297
>>>         msg = 0x7fc1e1c35360
>>>         ctx = {rec_lev = 0, run_flags = 0, last_retcode = 1, jmp_env 
>>> = {{__jmpbuf = {1048575, -3596212518023615478, 140470693039152,
>>>                 140470299422208, 140470299422208, 140733262279292, 
>>> 140733262279292, 140733262279284}, __mask_was_saved = -900960744,
>>>               __saved_mask = {__val = {5406222, 532575944923, 
>>> 541165879417, 4294967297, 1042, 140470299422952, 5409666, 65535,
>>>                   140733262279284, 140470300510264, 140470697171281, 
>>> 1042, 140470300510984, 18446744072809678880, 16, 17179869210}}}}}
>>>         ret = <value optimized out>
>>>         inb = {
>>>           s = 0x23d7cc0 "OPTIONS 
>>> sip:gavin.llewellyn at crocodiletalk.com SIP/2.0\r\nVia: SIP/2.0/TCP 
>>> edge00-int.crocodilertc.net:5080;branch=z9hG4bKab92.bb8249afcf13f20080f25121e49865b8.0\r\nVia: 
>>> SIP/2.0/WSS qvis2mie4gas.invalid;rp"..., len = 1028}
>>>         __FUNCTION__ = "receive_msg"
>>> #5  0x000000000052a251 in tcp_read_req (con=0x7fc1ca4c6e00, 
>>> bytes_read=0x7fff041b327c, read_flags=0x7fff041b3274) at 
>>> tcp_read.c:1387
>>>         bytes = <value optimized out>
>>>         total_bytes = 1028
>>>         resp = 1
>>>         size = <value optimized out>
>>>         req = 0x7fc1ca4c6e80
>>>         dst = {send_sock = 0x14, to = {s = {sa_family = 1,
>>>               sa_data = "\000\000\001\000\000\000\001 
>>> \000\000x\313\306", <incomplete sequence \341>}, sin = {sin_family = 
>>> 1, sin_port = 0,
>>>               sin_addr = {s_addr = 1}, sin_zero = "\001 
>>> \000\000x\313\306", <incomplete sequence \341>}, sin6 = {sin6_family 
>>> = 1,
>>>               sin6_port = 0, sin6_flowinfo = 1, sin6_addr = {__in6_u 
>>> = {
>>>                   __u6_addr8 = "\001 
>>> \000\000x\313\306\341\301\177\000\000\000\000\000", __u6_addr16 = 
>>> {8193, 0, 52088, 57798, 32705, 0, 0,
>>>                     0}, __u6_addr32 = {8193, 3787901816, 32705, 
>>> 0}}}, sin6_scope_id = 68891240}}, id = 32767, proto = 8 '\b', 
>>> send_flags = {
>>>             f = 0 '\000', blst_imask = 0 '\000'}}
>>>         c = 13 '\r'
>>>         ret = <value optimized out>
>>>         __FUNCTION__ = "tcp_read_req"
>>> #6  0x000000000052c53b in handle_io (fm=<value optimized out>, 
>>> events=1, idx=-1) at tcp_read.c:1617
>>>         ret = <value optimized out>
>>>         n = <value optimized out>
>>>         read_flags = 1
>>>         con = 0x7fc1ca4c6e00
>>>         s = <value optimized out>
>>>         resp = <value optimized out>
>>>         t = <value optimized out>
>>>         __FUNCTION__ = "handle_io"
>>> #7  0x000000000052eb69 in io_wait_loop_epoll (unix_sock=<value 
>>> optimized out>) at io_wait.h:1092
>>>
>>>
>>
>
>

-- 
Daniel-Constantin Mierla - http://www.asipto.com
http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda
Kamailio Advanced Trainings - Berlin, Nov 25-28; Miami, Nov 18-20, 2013
   - more details about Kamailio trainings at http://www.asipto.com -




More information about the sr-dev mailing list